Graph related

Type of graph

PlotS presents users with a range of eight graph types to select from:

Certain graph like density, frequency and histogram require only X-axis. The variable for Y-axis for the other remaining graph has to be numeric variable. Users can interactively change the variable for the axes.


Aesthetic option

The aesthetic choice serves as a valuable function that links a variable to a visual element like color, shape, or line type (dash, dotted, solid). This enables users to add additional variables or differentiate between variables. This functionality equips PlotS to effectively manage a wide range of data variables for analysis, setting it apart from other visualization tools.

Change color of variables:

  1. auto filled - by default PlotS will automatically select color for the chosen variable.

  2. customize - users can edit the color. They can either type the color in the text box (comma or space separated) or choose from the color picker provided at the side.

To illustrate Aesthetic options, we will use a hypothetical gene expression dataset (refer to Table 1) representing two rice cultivars (IR64 and N22) exposed to two types of treatments (t1 and t2), along with a control (c). Each condition has two replicates (R1, R2). Let’s create a scatter plot with aesthetic color (auto filled) mapped to treatment and Shape to replicate of the data. The resulting graphical representation is depicted in Figure 1.

Table 1. Expression data with two replicates of two rice cultivars under different treatment conditions.

cultivar

treatment

replicate

fpkm

IR64

t1

R1

20.90

IR64

t1

R2

17.75

IR64

t2

R1

5.90

IR64

t2

R2

3.39

IR64

c

R1

7.60

IR64

c

R2

6.60

N22

t1

R1

10.37

N22

t1

R2

11.93

N22

t2

R1

41.51

N22

t2

R2

33.64

N22

c

R1

23.81

N22

c

R2

28.01

Aesthetic setting
Aesthetic setting
Figure 1. Scatter plot with the chosen aesthetic elements - color and shape

Figure 1. Scatter plot with the chosen aesthetic elements - color and shape


Visualization of multivariate data

PlotS offers various features for multivariate analysis in addition to the features provided under Aesthetic options. Visualization of the relationship of multiple variables in a data can be done in four ways:

  1. Faceting

  2. Secondary Y-axis

  3. Side graph

  4. Inset

Faceting

Faceting generate sets of visual representations by partitioning data into smaller groups and showcasing identical graphs for each subgroup.

User has to select the Facet type. There are two types:

  1. wrap

  2. grid

Users can specify the number of column and row for facet. The value “0” signifies the default setting.

To exemplify the feature of faceting, we will utilize the provided Table 2. Although the data follows a format akin to Table 1, it contains a more comprehensive range of rows, providing a more detailed perspective.

Facet wrap setting
Facet wrap setting
Figure 2. displaying the wrap faceting

Figure 2. displaying the wrap faceting

Facet grid setting
Facet grid setting
Figure 2. displaying the grid faceting

Figure 2. displaying the grid faceting

Secondary Y-axis

Secondary Y-axis is to display two sets of data on the same graph, each with its own Y-axis and scale. It is useful for comparing two different types of data that have distinct measurement ranges, or when you want to show relationships between variables that might not be immediately apparent on a single Y-axis

To add the secondary Y-axis:

  1. Data must have two numeric variables (or columns)

  2. Users have to select any one graph listed under the Add layer with secondary Y-axis

Secondary Y-axis
Secondary Y-axis

Side graph

The Side graph is used to add graphs on the right side (Y-graph) and/or the upper side (X-graph) of the primary graph. This functionality within PlotS offers an additional and valuable avenue for investigating the relationship between variables. This feature is conveniently located beneath the main graph panel (shown below).

By default, the variables selected for the side graphs are aligned with those assigned to the X-axis and Y-axis of the main graph. However, users have the autonomy to opt for different variables when generating these side graphs.

It’s important to note that the Side graph feature is not compatible with the two-way ANOVA or when the secondary Y-axis is in use.

Side graph setting
Side graph setting

Inset

Inset is to insert a graph within the primary graph. The feature enhances the exploration and presentation of data by enabling users to focus on specific data points while still maintaining an overall view of the dataset. Inset panel is also placed below the primary graph panel along with the Side graph

  1. Add inset: Users can easily add an inset graph by using the mouse to select and drag across a specific data point of interest on the primary graph. The corresponding data from that point is then displayed below the panel. This data can also be downloaded as a CSV file.

  2. Add metadata: The variable displayed in the inset graph can be different from the one in the primary graph. This allows users to incorporate additional contextual information or metadata into the graph.

  3. Toggle Inset: By default, the Inset is turned on, but users can disable it if desired. When deactivated, the Inset will only turn off the display of the inset graph.

Inset of histogram
Inset of histogram

Additional layer without secondary y-axis

It add additional layer over the existing graph, but without secondary y-axis. It enable user to look for pattern in the data.

Type of layer:

  1. line graph

  2. point graph

  3. smooth line graph: applicable only when continuous data is provided to X-axis

Methods for plottig smooth line

  1. Linear regression model (LM)

  2. Generalized LM

  3. Generalized additive model

  4. LOESS : applicable for data having less than 1000 observations

Smooth line
Smooth line

Add error bar to a graph

Based on the graph, users have the option to incorporate a measure of variability using the Add error bar function.

There are different types of measurements available for this purpose:

  1. Confidence Interval at 95% Level: This measurement provides a range within which the true population parameter is likely to lie with a 95% confidence.

  2. Standard Error: This measurement represents the average variability between sample means and the actual population mean.

  3. Standard Deviation: This measurement quantifies the dispersion or spread of data points around the mean.

To add error bars to a graph, there are two approaches:

  1. Specifying Computed Statistics: Users can directly input the computed statistic values from a column in their dataset. This allows for precise control over the error bar representation.

  2. Auto Compute: Alternatively, PlotS offers an automatic computation option. It calculates the dispersion based on the data.

These error bars enhance the graphical representation by providing insights into the variability and reliability of the data points, aiding in a more comprehensive interpretation of the information presented.

Error bar feature
Error bar feature